MicroRNA Signature of Malignant Mesothelioma with Potential Diagnostic and Prognostic Implications Sara Busacca1*, Serena Germano1*, Loris De Cecco2,3, Maurizio Rinaldi1, Federico Comoglio1, Francesco Favero4, Bruno Murer5, Luciano Mutti6, Marco Pierotti2,3, and Giovanni Gaudino1 1 Dipartimento di Scienze Chimiche, Alimentari, Farmaceutiche e Farmacologiche, University of Piemonte Orientale, Novara, Italy; 2Molecular Cancer Genetics Group, Istituto FIRC di Oncologia Molecolare, Milan, Italy; 3Scientific Directorate Fondazione Istituto di Ricerca e Cura a Carattere Scientifico, Istituto Nazionale Tumori, Milan, Italy; 4Laboratory of Cancer Genomics, Fondo Edo Tempia, Biella, Italy; 5Department of Anatomic Pathology, Dell’Angelo Hospital, Zelarino, Italy; and 6Department of Medicine, Local Health Unit 11, Borgosesia, Italy

MicroRNAs (miRNAs) post-transcriptionally regulate the expression of target genes, and may behave as oncogenes or tumor suppressors. Human malignant mesothelioma is an asbestos-related cancer, with poor prognosis and low median survival. Here we report, for the first time, a cross-evaluation of miRNA expression in mesothelioma (MPP-89, REN) and human mesothelial cells (HMC–telomerase reverse transcriptase). Microarray profiling, confirmed by real-time quantitative RT-PCR, revealed a differential expression of miRNAs between mesothelioma and mesothelial cells. In addition, a computational analysis combining miRNA and gene expression profiles allowed the accurate prediction of genes potentially targeted by dysregulated miRNAs. Several predicted genes belong to terms of Gene Ontology (GO) that are associated with the development and progression of mesothelioma. This suggests that miRNAs may be key players in mesothelioma oncogenesis. We further investigated miRNA expression on a panel of 24 mesothelioma specimens, representative of the three histotypes (epithelioid, biphasic, and sarcomatoid), by quantitative RT-PCR. The expression of miR-17–5p, miR-21, miR-29a, miR-30c, miR-30e–5p, miR-106a, and miR-143 was significantly associated with the histopathological subtypes. Notably, the reduced expression of two miRNAs (miR-17–5p and miR-30c) correlated with better survival of patients with sarcomatoid subtype. Our preliminary analysis points at miRNAs as potential diagnostic and prognostic markers of mesothelioma, and suggests novel tools for the therapy of this malignancy. Keywords: human malignant mesothelioma; microRNAs; microarray; real-time quantitative RT-PCR; survival

Malignant mesothelioma is an aggressive cancer refractory to current therapies, the incidence of which is expected to rise in the next decades (1). Exposure to asbestos is a well known risk factor, although activated oncogenes, genetic predisposition, and SV40 infection were also proposed as cofactors (2). The

(Received in original form February 15, 2009 and in final form April 29, 2009) * S.B. and S.G. contributed equally to this work. This work was supported by the Associazione Italiana per la Ricerca sul Cancro, the Buzzi Unicem Foundation, Regione Piemonte RSF (G.G.), and GIMe (Mesothelioma Italian Group) (G.G.). This work includes experiments using array-based technology: array designs and experiments were submitted to ArrayExpress. For microRNA array, the MIAME (Minimum Information about a Microarray Experiment) accession numbers are: A-MEXP-1260 and E-MEXP-1699. For cDNA arrays, the MIAME accession numbers are: A-MEXP-1261, A-MEXP-1262, E-MEXP-1700, and E-MEXP-1701. Correspondence and requests for reprints should be addressed to Giovanni Gaudino, Ph.D., Dipartimento di Scienze Chimiche, Alimentari, Farmaceutiche e Farmacologiche, Via Bovio 6, 28,100 Novara, Italy. E-mail: ggaudino@crch. hawaii.edu This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org Am J Respir Cell Mol Biol Vol 42. pp 312–319, 2010 Originally Published in Press as DOI: 10.1165/rcmb.2009-0060OC on June 5, 2009 Internet address: www.atsjournals.org

CLINICAL RELEVANCE This work reveals that several microRNAs (miRNAs) are differentially expressed in malignant mesothelioma and in mesothelial cells. Some of their putative target genes are associated with the etiology of malignant mesothelioma. This study demonstrates dysregulated miRNAs as potential diagnostic biomarkers and as prognostic factors for malignant mesothelioma.

mechanism of asbestos carcinogenesis has been linked to the activation of proinflammatory cytokines and NF-kB (3, 4). Moreover, we previously demonstrated that malignant transformation of mesothelial cells upon exposure to asbestos and asbestos-like fibers occurs also through the activation of Akt (5, 6). Three histopathologic subtypes of mesothelioma were described—epithelioid, biphasic, and sarcomatoid—associated with different prognoses (7). Neither effective therapies, nor suitable prognostic tools, are currently available for malignant mesothelioma. Therefore, the knowledge of major molecular pathways involved in mesothelioma oncogenesis is necessary to define appropriate targets for the treatment of this disease. One of the most recent advances in cancer oncogenesis involves miRNAs, 22-nucleotide-long single-stranded RNAs that negatively regulate gene expression by modulating translational efficiency of target mRNAs. miRNAs are involved in numerous cellular processes, including development, differentiation, proliferation, apoptosis, and stress response (8). They may also function as oncogenes or tumor suppressor genes, and miRNAs aberrantly expressed have been described in several neoplasms, including breast, colon, prostate, and lung cancers (9, 10). Recent data indicate that miRNAs may be diagnostic markers for cancers and could correlate closely with patient survival and with several clinical–pathologic factors (11). The identification of a signature of miRNA expression between tumor cells and their normal counterparts is important to investigate the specific roles of the various miRNAs in tumor development, and to identify the genes and pathways that they target. Moreover, analysis on tumor tissues can provide new biomarkers for diagnosis and clinical outcome of this cancer. Nevertheless, the expression of miRNAs in mesothelioma cells has not yet been fully investigated. Here, we report that specific miRNAs are differentially expressed in mesothelioma and mesothelial cells. Algorithms for the prediction of putative miRNA targets were exploited in the frame of a gene expression profile of mesothelioma cells, revealing a number of genes previously implicated in the onset and progression of this malignancy. Furthermore, a survey on 24 mesothelioma tissue specimens of the three histological subtypes revealed that the expression of particular miRNAs

Busacca, Germano, De Cecco, et al.: miRNA Profiling in Human Malignant Mesothelioma

was altered in malignant mesothelioma, suggesting a correlation with patient survival in the sarcomatoid subtype.

MATERIALS AND METHODS Cell Lines and Tissues Human mesothelioma cells, MPP-89 (12), from the IST Cell Depository (Genoa, Italy), and REN (13), kindly provided by Dr. S. M. Albelda (University of Pennsylvania, Philadelphia, PA), were grown in their appropriate conditions: nutrient mixture F12 Ham (Sigma, St. Louis, MO) and 10% FBS (Invitrogen, Carlsbad, CA). Immortalized human mesothelial cells (HMC–telomerase reverse transcriptase [TERT]) (14) were grown in Medium 199, 10% FBS, 3.3 nM epidermal growth factor, 400 nM hydrocortisone, and 570 nM insulin. MPP-89 cells were characterized as hyperdiploid, bearing deletions/translocations of chromosome 6, and chromosome 13 monosomy. Moreover, they lack both chromosomes 17. REN cells contain a rearranged p53 gene, and lack expression of p53 protein (15), whereas no abnormalities were reported for HMC-TERT (14). The 24 primary tumor tissue specimens were from the Department of Anatomic Pathology, Dell’Angelo Hospital (Zelarino–Venice, Italy). The tissue sections were representative of the three hystotypes: eight epithelioid, eight biphasic, and eight sarcomatoid. These sections were highly cellular, with a large neoplastic component (.80%), and were collected at diagnosis (patients did not receive previous treatments). Ethical background: studies on tissue specimens do not require the approval by an ethics committee, according to Italian national rules.

RNA Isolation Total RNA was extracted from cells by standard procedure with Trizol (Invitrogen), and from tissue sections by FFPE RNeasy Kit (Qiagen, Valencia, CA), according to manufacturers’ instructions. RNA concentration was determined with Quant-iT RNA Fluorimetric Assay Kit (Invitrogen), and integrity was confirmed by gel electrophoresis.

miRNAs Microarray Analysis miRNA expression profiling was performed with the LNA-modified oligonucleotide microarrays (miRCURY locked nucleic acid [LNA] microarray kit, Version 8.0; Exiqon, Vedbaek, Denmark). Total RNA (8 mg) was labeled with a fluorescent dye (Hy3, Hy5) following the manufacturer’s instructions. The samples were hybridized for 16 hours at 608C in a hybridization station (Genomic Solutions, Ann Arbor, MI). Slides were washed using buffer A at 608C, then buffer B and C at room temperature, and dried by centrifugation. Slides were scanned using the GenePix 4000B microarray scanner (Axon Instruments, Union City, CA). A ‘‘dye swap’’ (fluorochrome inversion) experimental design was followed to avoid biases due to different properties of the labels. Three biological replicates were performed. MIAME (Minimum Information about a Microarray Experiment) accession numbers are A-MEXP-1260 and E-MEXP-1699.

Real-Time Quantitative RT-PCR Reverse transcription was performed with TaqMan MicroRNA Reverse Transcription kit (Applied Biosystem, Foster City, CA) using specific miRNA primers, or RevertAid H Minus First Strand cDNA Synthesis kit (Fermentas, Burlington, Ontario, Canada). Reaction was performed on 50 ng or 1 mg of total RNA. Real-time PCR was performed using TaqMan Gene Expression Master Mix, TaqMan MicroRNA assays (Applied Biosystems) or TaqMan gene expression assays (Applied Biosystems). Reactions were incubated for 10 minutes at 958C, followed by 40 cycles of 15 seconds at 958C and 1 minute at 608C. Analysis of relative miRNA and mRNA expression were performed using the DDCT method with U6 or glyceraldehyde 3-phosphate dehydrogenase as endogenous controls, respectively. The expression of 10 miRNAs was analyzed in 24 mesothelioma specimens by quantitative RT-PCR (qRT-PCR). A calibrator sample was prepared by pooling an equal amount of RNA from each specimen, and miRNA expression levels are expressed as a relative fold increase over this common reference, allowing a quantitative comparison of the different specimens.

313

cDNA Microarrays Gene expression analysis was performed as previously described (16). Briefly, cDNA fragments were spotted in triplicate on UltraGAPS amino-silane–coated slides (Corning, Life Sciences, Amsterdam, The Netherlands), using the Lucidea Microarray Spotter (Amersham Biosciences, Amersham, UK). The target cDNAs were synthesized from total RNA and labeled with Cy3-dCTP (reference RNA) or Cy5-dCTP (sample RNA) (Amersham Biosciences), and indirectly with 3DNA Array Detection 900 kit (Genisphere, Montvale, NJ). Total RNA was reverse transcribed using 50 end-modified oligo-dT primers containing the specific Cy3 or Cy5 3DNA capture sequences. The 32P-labeled cDNAs were annealed with specific 3DNA capture reagent for 16 hours at 41 C8. Hybridization was performed in a hybridization station, and slides were scanned using the GenePix 4000B microarray scanner. MIAME accession numbers are A-MEXP-1261, A-MEXP-1262, EMEXP-1700, and E-MEXP-1701.

Bioinformatics and Statistical Analysis Microarrays data were analyzed using R version 2.7.2 (a free language and environment for statistical computing; R Development Core Team 2005, www.R-project.org). Normexp background correction and global loess normalization were performed using the Limma package (17). Differentially expressed probes and the relevant statistics were detected using eBayes command in Limma. Gene set enrichment analysis (GSEA) was performed to identify the biological patterns of the genes analyzed for gene expression (18). Each gene was associated with the corresponding Gene Ontology (GO) terms, according to the GO annotation for Biological Process. For any given GO term, K, the index, zK, was calculated according to the formula: 1 zK 5 pffiffiffiffiffiffi + tk nK k2K where k ranges over all genes of the given GO term, K, tk is the value of the Student’s t value for the comparison, and nK is the number of genes of the GO term. We selected GO terms with 3 < nK < 100. The significance threshold for the permutation test was set at a P value of 0.05. Predicted targets of miRNAs were identified by using four databases: PicTar (19), TargetScan (20), MiRanda (21), and Diana-microT (22). Data obtained from the prediction analysis were filtered by matching with those from the gene expression analysis. For each correlation between miRNAs and gene, the occurrences in the four databases have been reported in a contingency table. The GSEA significant GO term recurrence in the contingency table was also investigated. Kruskal–Wallis test on the 24 patients with mesothelioma was performed. Patient survival curves were estimated by the Kaplan– Meier method and Cox proportional hazards regression model.

RESULTS miRNA Profiling of Mesothelioma Cells

We used a microarray approach to analyze miRNA profiles of two mesothelioma cell lines (MPP-89 and REN) representative of two distinct morphological types (i.e., ‘‘spindle-shaped’’ and ‘‘epithelial-like’’ cells, respectively) and of HMC-TERT nontumor mesothelial cells. In the three biological replicates performed, we consistently identified a total of 65 dysregulated miRNAs (Figure 1A). Both mesothelioma cells shared the same 10 up-regulated and 19 down-regulated miRNAs, when compared with HMC-TERT, whereas a higher number of miRNAs were up-regulated in MPP-89 than in REN cells (Figure 1B). Microarray data were validated by qRT-PCR on 10 miRNAs (Figure 1C). The differences in miRNA expression between mesothelioma and mesothelial cells were consistent, except for one single discrepancy observed for hsa-miR-21. When com-

314

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 42 2010

Figure 1. miRNA expression profile in mesothelioma cells. (A) Significant differences (P < 0.05, Student’s t test of three biological replicates) in the paired comparisons of human mesothelial cells (HMC)-TERT, MPP-89, and REN samples are reported in a Venn diagram. Intersections represent differentially expressed miRNAs in either comparison. (B) Cross-evaluation of microRNA (miRNA) expression (red, higher; green, lower) in three different cell lines: MPP-89 and REN (mesothelioma cells) compared with HMC-TERT (nontumoral cells). The analysis also allowed the evaluation of miRNAs differentially modulated between the two tumor cell lines. #miRNA significantly modulated when hybridized with two different probes. (C) Expression analysis by quantitative RT-PCR (qRT-PCR) of 10 selected miRNAs for microarray data validation. Data are expressed as relative fold change over the value of HMC-TERT.

paring MPP-89 with REN cells, similar changes were obtained by both assays, with the exception of hsa-miR-30e–5p. Gene Expression and Biological Significance of miRNA Dysregulation in Mesothelioma

The effects of dysregulated miRNAs on the onset and progression of malignant mesothelioma can be reflected by the expression changes of their putative target genes. To identify the effective regulatory activities of miRNAs, we integrated miRNA and gene expression microarray data with miRNA target predictions.

To map putative targets of mesothelioma dysregulated miRNAs, we first exploited the four current algorithms (miRanda, TargetScan, PicTar, and Diana-microT); afterward, we filtered data from the prediction analysis by matching with those obtained from gene expression profile of MPP-89, REN, and HMC-TERT cells. The gene expression profile was performed by a cDNA microarray, and this analysis revealed that 3,277 genes were modulated in MPP-89 and REN cells, with a positive correlation (r 5 10.725) for gene up- or down-regulation, in both mesothelioma cells, when compared with mesothelial cells (Figure 2A). We validated the expression of 10 genes by qRTPCR (see Table E6 in the online supplement), selected by one

Busacca, Germano, De Cecco, et al.: miRNA Profiling in Human Malignant Mesothelioma

315

Figure 2. Gene expression profile and target prediction. (A) Scatter plot of Log fold changes of differentially expressed genes in MPP-89 and REN cells compared with HMC-TERT cells. The top 50 genes significantly modulated are colored in a red to dark brown ladder representing the Logadjusted P value. (B) miRNA/genes contingency table: data from the prediction analysis were matched with those from gene expression analysis. On the x axis is indicated the name of the first of every fifteen genes out of the 587 identified among modulated genes. On the y axis are reported 46 out of the 65 modulated miRNAs, for which at least one target was identified among the gene products. Dark blue lines indicate target genes predicted by all four algorithms. Progressive decrease in blue intensities indicates target genes predicted by three, two, or one algorithm. The detailed list of all genes is reported in the original matrix (Table S4).

of the following two criteria: (1) maximum number of algorithms predicting the same miRNAs targeting the gene; and (2) high selectivity of miRNA targeting the gene, irrespective of the number of algorithms. The combination of data from miRNA and gene expression profiles with target prediction was instrumental in generating a contingency table (Figure 2B, Table E1), in which 587 genes are represented, identified as putative targets of dysregulated miRNAs. To identify the signaling pathways related to the genes found differentially expressed in mesothelioma cells and putative targets of dysregulated miRNAs, we performed a Gene Ontol-

ogy analysis. Several differentially regulated GO terms were identified when mesothelioma cells were compared with mesothelial cells (Tables E2 and E3), and when mesothelioma cells were compared with each other (Table E4). P-values of up-regulated genes are highlighted in red and those of downregulated genes are highlighted in green in these tables. Interestingly, pathways controlling programmed cell death (mitochondrial release of cytochrome c, regulation of caspase activity), chronic inflammatory response, NO-mediated signaling, cell growth and motility, and protein kinase activity, frequently associated with mesothelioma carcinogenesis (1, 23, 24), were found among the significant GO terms.

316

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 42 2010

We then investigated the recurrence in the contingency table of the GO terms that were found significant in the GSEA performed (Table E5). miRNAs May Be Diagnostic and Prognostic Markers for Malignant Mesothelioma

Dysregulation of miRNAs expression has been associated with tissue specificity, differentiation, and patient survival for different cancers (11, 25). Our results on cultured cells prompted us to verify whether the expression of specific miRNAs may correlate with various biopathologic features associated with tumor specimens, and thus may represent diagnostic markers in mesothelioma tissues. Here, we performed a survey of 24 paraffin-embedded tissue specimens, consisting of eight epithelioid, eight biphasic, and eight sarcomatoid mesotheliomas. The patients were characterized by their age, sex, survival, and tumor stage (Table 1). Ten miRNAs, selected from the contingency table and analyzed by qRT-PCR, were found expressed in all tumor tissues to different extents. The expression of 7 out of 10 miRNAs analyzed was significantly associated with the histopathological subtypes. In particular, differences were observed for subtype pairs biphasic–sarcomatoid and epithelioid– sarcomatoid, with the exception of hsa-miR-30c, which is differentially expressed in all three subtypes (Table E7 and Figure 3). No association with the histopathologic subtypes was found for hsa-miR-31, hsa-miR-221, or hsa-miR-222. Moreover, we verified whether miRNA expression could be a possible indicator of survival. Patient samples of the same subtype were divided in two groups, above or below the median of expression calculated for each miRNA, and Kaplan–Meier curves were obtained. In epithelioid and biphasic subtypes, no correlations were observed between survival and miRNA expression. In contrast, for sarcomatoid mesotheliomas, the lower expression of hsa-miR-17–5p and hsa-miR-30c was significantly associated with better survival (Figure 4). Interestingly, for these two miRNAs, all the patients with expression values below the median (light gray) had a longer survival than

the others (dark gray). However, no correlations were observed with tumor stage. The number of specimens tested here is quite low; however, a European Organization for Research and Treatment of Cancer (EORTC) prognostic score was calculated for each patient of the sarcomatoid subtype as the combination of five parameters: age, sex, histology, probability of diagnosis, and leukocyte count (26). According to the EORTC prognostic score system, all patients were stratified in the high-risk group, and this indicates that the observed association between survival and lower expression of the two miRNAs is remarkably relevant.

DISCUSSION Our miRNA expression profile on mesothelioma and mesothelial cells revealed that miRNAs linked to oncogenesis, such as hsa-miR-17–3p, hsa-miR-17–5p, hsa-miR-18a, and hsa-miR-20a, belonging to the miR-17–92 cluster (27), were up-regulated in mesothelioma cells. In addition, hsa-miR-21, hsa-miR-29a, hsamiR-30b, and hsa-miR-106a were significantly dysregulated, in accordance with results previously reported for other solid tumors, including lung cancer (10). On the other hand, we observed that hsa-miR-31 was markedly down-regulated in mesothelioma cells, in contrast to overexpression observed in colorectal carcinoma cells (28). Similarly, hsa-miR-143 was strongly overexpressed in MMP-89 cells, whereas it is down-regulated in colorectal carcinoma cells (25). In REN cells, hsa-miR-143 was also up-regulated, but to a markedly lesser extent. In addition, hsa-miR-30c was upregulated in mesothelioma cells, the only miRNA expressed more in REN than in MPP-89. Finally, hsa-miR-221 and hsamiR-222 were equally down-regulated in both mesothelioma

TABLE 1. CLINICAL FEATURES OF MALIGNANT MESOTHELIOMA SPECIMENS Histotype Epithelioid Epithelioid Epithelioid Epithelioid Epithelioid Epithelioid Epithelioid Epithelioid Biphasic Biphasic Biphasic Biphasic Biphasic Biphasic Biphasic Biphasic Sarcomatoid Sarcomatoid Sarcomatoid Sarcomatoid Sarcomatoid Sarcomatoid Sarcomatoid Sarcomatoid

Age (yr)

Sex

70 67 54 57 56 67 58 49 74 60 70 60 70 74 60 73 53 56 61 64 56 67 67 64

Male Male Male Male Female Female Male Male Male Male Male Male Male Female Female Male Male Male Male Male Male Male Male Male

Stage T3 T4 T3 T4 T2 T2 T3 T4 T2 T2 T3 T2 T4 T4 T3 T3 T4 T4 T4 T3 T3 T3 T3 T2

N0 N0 N1 N0 N0 N0 N1 N0 N0 N0 N0 N0 N1 N0 N0 N1 N0 N0 N0 N0 N1 N1 N0 N0

M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 M0

Survival (mo) 24 9 4 7 73 18 48 28 10 44 11 12 2 7 19 11 1 8 4 18 5 10 1 3

Definition of abbreviations: M, distant metastasis; N, lymph nodes involved; T, size and staging of the tumor.

Figure 3. miRNA expression in mesothelioma specimens. miRNAs expressed in mesothelioma tissue specimens, representative of the three phenotypes, were analyzed by qRT-PCR. The box plot shows the distribution of expression data quartiles for each miRNA and for each histotype. Data are expressed as relative fold increase over a representative common calibrator, obtained by pooling the RNAs of all samples examined. The black horizontal line of each box is the median. Minimum and maximum values are the bar end points.

Busacca, Germano, De Cecco, et al.: miRNA Profiling in Human Malignant Mesothelioma

317

Figure 4. miRNAs and patient survival. Survival curves related to miRNA expression in specimens from patients of the sarcomatoid subtype. Cases were stratified in two groups according to the median of each miRNA expression. Dark and light gray curves show patients with miRNA expression values above or below the median, respectively. P values for differences between curves are shown.

cells, as in breast cancer, whereas they are up-regulated in gastrointestinal tumors (10). The family of let-7 miRNAs has been shown to be downregulated in several tumors, suggesting a role as oncosuppressors. In both mesothelioma cells, several members of the hsa-let-7 family were moderately up-regulated instead. However, hsa-let-7a and other members are mediators of IL-6– dependent survival in human cholangiocytes (29), and IL-6 is active on mesothelioma cells, including MPP-89 (12). Therefore, one can hypothesize that let-7 overexpression may contribute to IL-6–induced survival in mesothelioma. The combination of target prediction with gene expression has been proposed as a useful tool to identify miRNA activity (30). Accordingly, we combined data obtained from the miRNA/mRNA microarray data with a computational analysis focused on target prediction. Our results revealed that some of the putative targets identified are genes associated with the etiology of malignant mesothelioma: cyclin-dependent kinase inhibitor 1B (CDKN1B), encoding p27Kip, correlates with survival (31); B cell translocation gene 1, shown as antiproliferative, has been found in human malignant pleural effusions (32); hepatocyte growth factor, shown by us as an autocrine growth

factor in mesothelioma cells (33); MECP2, a methyl-CpG– binding protein repressing transcription of genes the promoters of which are methylated, whereas CpG methylation has been reported as a tumorigenic mechanism of action of asbestos (34). Further studies will be performed to validate the functional role of the miRNAs targeting these genes, and to investigate the mechanisms of their interaction. Indeed, miRNAs are known to interact with the 39-untranslated region of target mRNAs; however, some miRNAs can also induce mRNA degradation (35). Interestingly, the contingency table represented in Figure 2B identified the down-regulated B cell translocation gene 1 oncosuppressor gene as the putative target of hsa-miR-106a, hsa-miR-17–5p and hsa-miR-143, which are all up-regulated, suggesting the possible involvement of this mechanism. To better understand the biological significance of dysregulated miRNAs and their putative targets in malignant mesothelioma, we performed a Gene Ontology analysis. Thus, we identified the most relevant GO terms involved in mesothelioma oncogenesis and possibly affected by dysregulated miRNAs. Dysregulation of miRNA expression has been associated with tissue specificity, differentiation, and patient survival for different cancers (11, 25). MiRNA expression profiling has been

318

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 42 2010

attempted for tumor classification on a panel of cancer specimens, including eight malignant mesotheliomas, exclusively of the epithelioid type (9). In view of these results, and on the basis of our miRNA expression profile in cultured cells, we verified whether expression analysis of miRNAs in mesothelioma tissues may suggest the occurrence of novel diagnostic markers. Interestingly, the expression was highest in epithelioid, intermediate in biphasic, and lowest in sarcomatoid subtypes, revealing an intriguing specificity of miRNA expression by the different histotypes. Although this work does not provide insights into mechanisms of oncogenesis, for which further validation studies will be required, our results identify dysregulated miRNAs as potential biomarkers for histology. The three different histopathologic subtypes are important prognostic factors, because epithelioid mesotheliomas have a more favorable prognosis than the sarcomatoid subtype, which is, in turn, more aggressive than the biphasic subtype (7). Therefore, we analyzed the possible correlation between miRNA expression and survival. Kaplan–Meier graphics reveal that such correlation exists for miR-17–5p and miR-30c only in mesotheliomas of the sarcomatoid subtype. This might be explained by the well established differences from the other two histotypes in aggressiveness and chemoresistance (7). Interestingly, hsa-miR-17–5p has been characterized as a powerful oncogene in neuroblastoma, where it sustains chemoresistance (36), whereas it behaves as an oncosuppressor in breast cancer (37), suggesting tissue specificity. Despite the small sample size (eight specimens), the highly significant P values (P 5 0.006) for log-rank analysis obtained here suggest that the observed correlation between expression of these miRNAs and patient survival may be not due to chance, and the effect of these miRNAs could be very large. Notably, the survival curves obtained for both miRNAs are clearly separated (e.g., the shortest survivors of the higher survival group died after the longest survivors of the lower survival group). It stands to reason that a well established picture of the possible prognostic significance of miR-17–5p and miR-30c in mesothelioma will require a further expansion of this study on a larger number of samples. Sarcomatoid mesothelioma is very rarely regarded as surgically operable, regardless of the stage of the disease. Our data address some underlying biological characteristics other than staging that might affect prognosis. Therefore, although the small study size does not allow us to be conclusive, our findings provide a genetic rationale for achieving a better definition of prognosis and of potential help in identifying operable patients belonging to the sarcomatoid subtype of malignant mesothelioma. Moreover, these data suggest miRNAs as potential tools for the development of novel therapeutic strategies for this cancer. Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Acknowledgments: The authors thank Dr. Giovanna Chiorino for helpful discussion and advice.

References 1. Pass HI, Vogelzang N, Hahn S, Carbone M. Malignant pleural mesothelioma. Curr Probl Cancer 2004;28:93–174. 2. Carbone M, Emri S, Dogan AU, Steele I, Tuncer M, Pass HI, Baris YI. A mesothelioma epidemic in cappadocia: scientific developments and unexpected social outcomes. Nat Rev Cancer 2007;7:147–154. 3. Yang H, Bocchetta M, Kroczynska B, Elmishad AG, Chen Y, Liu Z, Bubici C, Mossman BT, Pass HI, Testa JR, et al. TNF-alpha inhibits asbestos-induced cytotoxicity via a NF-kappaB–dependent pathway, a possible mechanism for asbestos-induced oncogenesis. Proc Natl Acad Sci USA 2006;103:10397–10402.

4. Sartore-Bianchi A, Gasparri F, Galvani A, Nici L, Darnowski JW, Barbone D, Fennell DA, Gaudino G, Porta C, Mutti L. Bortezomib inhibits nuclear factor–kappaB dependent survival and has potent in vivo activity in mesothelioma. Clin Cancer Res 2007;13:5942– 5951. 5. Cacciotti P, Barbone D, Porta C, Altomare DA, Testa JR, Mutti L, Gaudino G. Sv40-dependent Akt activity drives mesothelial cell transformation after asbestos exposure. Cancer Res 2005;65:5256–5262. 6. Bertino P, Marconi A, Palumbo L, Bruni BM, Barbone D, Germano S, Dogan AU, Tassi GF, Porta C, Mutti L, et al. Erionite and asbestos differently cause transformation of human mesothelial cells. Int J Cancer 2007;121:12–20. 7. Robinson BW, Lake RA. Advances in malignant mesothelioma. N Engl J Med 2005;353:1591–1603. 8. Miska EA. How microRNAs control cell division, differentiation and death. Curr Opin Genet Dev 2005;15:563–568. 9. Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D, SweetCordero A, Ebert BL, Mak RH, Ferrando AA, et al. MicroRNA expression profiles classify human cancers. Nature 2005;435:834–838. 10. Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F, Visone R, Iorio M, Roldo C, Ferracin M, et al. A microRNA expression signature of human solid tumors defines cancer gene targets. Proc Natl Acad Sci USA 2006;103:2257–2261. 11. Guo Y, Chen Z, Zhang L, Zhou F, Shi S, Feng X, Li B, Meng X, Ma X, Luo M, et al. Distinctive microRNA profiles relating to patient survival in esophageal squamous cell carcinoma. Cancer Res 2008;68:26–33. 12. Orengo AM, Spoletini L, Procopio A, Favoni RE, De Cupis A, Ardizzoni A, Castagneto B, Ribotta M, Betta PG, Ferrini S, et al. Establishment of four new mesothelioma cell lines: characterization by ultrastructural and immunophenotypic analysis. Eur Respir J 1999; 13:527–534. 13. Smythe WR, Kaiser LR, Hwang HC, Amin KM, Pilewski JM, Eck SJ, Wilson JM, Albelda SM. Successful adenovirus-mediated gene transfer in an in vivo model of human malignant mesothelioma. Ann Thorac Surg 1994;57:1395–1401. 14. Dickson MA, Hahn WC, Ino Y, Ronfard V, Wu JY, Weinberg RA, Louis DN, Li FP, Rheinwald JG. Human keratinocytes that express hTERT and also bypass a p16(ink4a)-enforced mechanism that limits life span become immortal yet retain normal growth and differentiation characteristics. Mol Cell Biol 2000;20:1436–1447. 15. Pietruska JR, Kane AB. Sv40 oncoproteins enhance asbestos-induced DNA double-strand breaks and abrogate senescence in murine mesothelial cells. Cancer Res 2007;67:3637–3645. 16. De Cecco L, Marchionni L, Gariboldi M, Reid JF, Lagonigro MS, Caramuta S, Ferrario C, Bussani E, Mezzanzanica D, Turatti F, et al. Gene expression profiling of advanced ovarian cancer: characterization of a molecular signature involving fibroblast growth factor 2. Oncogene 2004;23:8171–8183. 17. Rascoe PA, Cao X, Daniel JC, Miller SD, Smythe WR. Receptor tyrosine kinase and phosphoinositide-3 kinase signaling in malignant mesothelioma. J Thorac Cardiovasc Surg 2005;130:393–400. 18. Gentleman RMM, Huber W. Gene set enrichment analysis. In: Hahne F, Gentleman R, Falcon S, editors. Bioconductor case studies. New York: Springer; 2008. pp. 193–205. 19. Lall S, Grun D, Krek A, Chen K, Wang YL, Dewey CN, Sood P, Colombo T, Bray N, Macmenamin P, et al. A genome-wide map of conserved microRNA targets in c. Elegans. Curr Biol 2006;16:460–471. 20. Lewis BP, Shih IH, Jones-Rhoades MW, Bartel DP, Burge CB. Prediction of mammalian microRNA targets. Cell 2003;115:787–798. 21. John B, Enright AJ, Aravin A, Tuschl T, Sander C, Marks DS. Human microRNA targets. PLoS Biol 2004;2:e363. 22. Kiriakidou M, Nelson PT, Kouranov A, Fitziev P, Bouyioukos C, Mourelatos Z, Hatzigeorgiou A. A combined computational–experimental approach predicts human microRNA targets. Genes Dev 2004;18:1165–1178. 23. Marrogi A, Pass HI, Khan M, Metheny-Barlow LJ, Harris CC, Gerwin BI. Human mesothelioma samples overexpress both cyclooxygenase2 (COX-2) and inducible nitric oxide synthase (NOS2): in vitro antiproliferative effects of a COX-2 inhibitor. Cancer Res 2000;60: 3696–3700. 24. Fennell DA, Rudd RM. Defective core-apoptosis signalling in diffuse malignant pleural mesothelioma: opportunities for effective drug development. Lancet Oncol 2004;5:354–362. 25. Michael MZ, O’Connor SM, van Holst Pellekaan NG, Young GP, James RJ. Reduced accumulation of specific microRNAs in colorectal neoplasia. Mol Cancer Res 2003;1:882–891.

Busacca, Germano, De Cecco, et al.: miRNA Profiling in Human Malignant Mesothelioma 26. Fennell DA, Parmar A, Shamash J, Evans MT, Sheaff MT, Sylvester R, Dhaliwal K, Gower N, Steele J, Rudd R. Statistical validation of the EORTC prognostic model for malignant pleural mesothelioma based on three consecutive phase II trials. J Clin Oncol 2005;23:184–189. 27. Mendell JT. Miriad roles for the miR-17–92 cluster in development and disease. Cell 2008;133:217–222. 28. Bandres E, Cubedo E, Agirre X, Malumbres R, Zarate R, Ramirez N, Abajo A, Navarro A, Moreno I, Monzo M, et al. Identification by real-time PCR of 13 mature microRNAs differentially expressed in colorectal cancer and non-tumoral tissues. Mol Cancer 2006;5:29. 29. Meng F, Henson R, Wehbe-Janek H, Smith H, Ueno Y, Patel T. The microRNA let-7a modulates interleukin-6–dependent STAT-3 survival signaling in malignant human cholangiocytes. J Biol Chem 2007; 282:8256–8264. 30. Cheng C, Li LM. Inferring microRNA activities by combining gene expression with microRNA target prediction. PLoS ONE April 23, 2008;3(4):e1989. 31. Baldi A, Santini D, Vasaturo F, Santini M, Vicidomini G, Di Marino MP, Esposito V, Groeger AM, Liuzzi G, Vincenzi B, et al. Prognostic significance of cyclooxygenase-2 (COX-2) and expression of cell cycle inhibitors p21 and p27 in human pleural malignant mesothelioma. Thorax 2004;59:428–433. 32. Bard MP, Hegmans JP, Hemmes A, Luider TM, Willemsen R, Severijnen LA, van Meerbeeck JP, Burgers SA, Hoogsteden HC, Lambrecht BN.

33.

34.

35.

36.

37.

319

Proteomic analysis of exosomes isolated from human malignant pleural effusions. Am J Respir Cell Mol Biol 2004;31:114–121. Cacciotti P, Libener R, Betta P, Martini F, Porta C, Procopio A, Strizzi L, Penengo L, Tognon M, Mutti L, et al. Sv40 replication in human mesothelial cells induces HGF/MET receptor activation: a model for viral-related carcinogenesis of human malignant mesothelioma. Proc Natl Acad Sci USA 2001;98:12032–12037. Christensen BC, Godleski JJ, Marsit CJ, Houseman EA, LopezFagundo CY, Longacker JL, Bueno R, Sugarbaker DJ, Nelson HH, Kelsey KT. Asbestos exposure predicts cell cycle control gene promoter methylation in pleural mesothelioma. Carcinogenesis 2008;29: 1555–1559. Lim LP, Lau NC, Garrett-Engele P, Grimson A, Schelter JM, Castle J, Bartel DP, Linsley PS, Johnson JM. Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs. Nature 2005;433:769–773. Fontana L, Fiori ME, Albini S, Cifaldi L, Giovinazzi S, Forloni M, Boldrini R, Donfrancesco A, Federici V, Giacomini P, et al. AntagomiR-17–5p abolishes the growth of therapy-resistant neuroblastoma through p21 and bim. PLoS One May 21, 2008;3:e2236. Yu Z, Wang C, Wang M, Li Z, Casimiro MC, Liu M, Wu K, Whittle J, Ju X, Hyslop T, et al. A cyclin d1/microRNA 17/20 regulatory feedback loop in control of breast cancer cell proliferation. J Cell Biol 2008;182: 509–517.